doi:10.1152/ajpcell.00044.2002 284:535-546, 2003. First published Oct 16, 2002;Am J Physiol Cell Physiol

Keith Nehrke, Claire C. Quinn and Ted Begenisich channels in parotid acinar cells Molecular identification of Ca2+-activated K+

You might find this additional information useful...

38 articles, 21 of which you can access free at: This article cites http://ajpcell.physiology.org/cgi/content/full/284/2/C535#BIBL

13 other HighWire hosted articles, the first 5 are: This article has been cited by

  [PDF]  [Full Text]  [Abstract]

, March 1, 2008; 294 (3): C810-C819. Am J Physiol Cell PhysiolT. Nakamoto, V. G. Romanenko, A. Takahashi, T. Begenisich and J. E. Melvin

exocrine glandApical maxi-K (KCa1.1) channels mediate K+ secretion by the mouse submandibular 

[PDF]  [Full Text]  [Abstract], July 1, 2008; 88 (3): 1119-1182. Physiol Rev

D. Heitzmann and R. Warth Physiology and Pathophysiology of Potassium Channels in Gastrointestinal Epithelia

  [PDF]  [Full Text]  [Abstract]

, January 1, 2009; 50 (1): 194-202. Invest. Ophthalmol. Vis. Sci.M. Irnaten, R. C. Barry, B. Quill, A. F. Clark, B. J. P. Harvey and C. J. O'Brien

Human Lamina Cribrosa CellsActivation of Stretch-Activated Channels and Maxi-K+ Channels by Membrane Stress of 

[PDF]  [Full Text]  [Abstract], April 1, 2009; 296 (4): C878-C888. Am J Physiol Cell Physiol

V. G. Romanenko, K. S. Roser, J. E. Melvin and T. Begenisich channels

The role of cell cholesterol and the cytoskeleton in the interaction between IK1 and maxi-K 

[PDF]  [Full Text]  [Abstract], January 1, 2010; 90 (1): 113-178. Physiol Rev

S. Wray and T. Burdyga Sarcoplasmic Reticulum Function in Smooth Muscle

including high-resolution figures, can be found at: Updated information and services http://ajpcell.physiology.org/cgi/content/full/284/2/C535

can be found at: AJP - Cell Physiologyabout Additional material and information http://www.the-aps.org/publications/ajpcell

This information is current as of July 8, 2010 .  

http://www.the-aps.org/.American Physiological Society. ISSN: 0363-6143, ESSN: 1522-1563. Visit our website at a year (monthly) by the American Physiological Society, 9650 Rockville Pike, Bethesda MD 20814-3991. Copyright © 2003 by the

is dedicated to innovative approaches to the study of cell and molecular physiology. It is published 12 timesAJP - Cell Physiology

on July 8, 2010 ajpcell.physiology.org

Dow

nloaded from

Molecular identification of Ca2�-activated K� channelsin parotid acinar cells

KEITH NEHRKE,1,* CLAIRE C. QUINN,1,* AND TED BEGENISICH2

1Center for Oral Biology, Aab Institute of Biomedical Sciences, and 2Department of Pharmacologyand Physiology, University of Rochester Medical Center, Rochester, New York 14642Submitted 29 January 2002; accepted in final form 15 October 2002

Nehrke, Keith, Claire C. Quinn, and Ted Begenisich.Molecular identification of Ca2�-activated K� channels inparotid acinar cells. Am J Physiol Cell Physiol 284:C535–C546, 2003. First published October 16, 2002; 10.1152/ajpcell.00044.2002.—We used molecular biological andpatch-clamp techniques to identify the Ca2�-activated K�

channel genes in mouse parotid acinar cells. Two types of K�

channels were activated by intracellular Ca2� with single-channel conductance values of 22 and 140 pS (in 135 mMexternal K�), consistent with the intermediate and maxi-Kclasses of Ca2�-activated K� channels, typified by the mIK1(Kcnn4) and mSlo (Kcnma1) genes, respectively. The pres-ence of mIK1 mRNA was established in acinar cells by in situhybridization. The electrophysiological and pharmacologicalproperties of heterologously expressed mIK1 channelsmatched those of the native current; thus the native, smallerconductance channel is likely derived from the mIK1 gene.We found that parotid acinar cells express a single, uncom-mon splice variant of the mSlo gene and that heterologouslyexpressed channels of this Slo variant had a single-channelconductance indistinguishable from that of the native, large-conductance channel. However, the sensitivity of this ex-pressed Slo variant to the scorpion toxin iberiotoxin wasconsiderably different from that of the native current. RT-PCR analysis revealed the presence of two mSlo �-subunits(Kcnmb1 and Kcnmb4) in parotid tissue. Comparison of theiberiotoxin sensitivity of the native current with that ofparotid mSlo expressed with each �-subunit in isolation andmeasurements of the iberiotoxin sensitivity of currents incells from �1 knockout mice suggest that parotid acinar cellscontain approximately equal numbers of homotetramericchannel proteins from the parotid variant of the Slo gene andheteromeric proteins composed of the parotid Slo variant incombination with the �4-subunit.

secretory cells; fluid secretion; single channels; patch clamp

SALIVARY GLAND SECRETION plays an extremely importantrole in maintaining the health of oral tissues. The fluidsecretion hydrates the oral cavity, aiding in the mas-tication and swallowing of food. In addition, thesesecretions neutralize acids and protect against theinvasion of potential pathogens. An increase in intra-cellular Ca2�, usually associated with muscarinic re-ceptor stimulation, triggers fluid secretion by simulta-

neously activating apical Cl� channels and basolateralK� channels. Indeed, nonselective K� channel blockersinhibit fluid secretion and Cl� efflux (19, 21). The effluxof Cl� and K� across the apical and basolateral mem-branes, respectively, produces a transepithelial poten-tial difference that draws Na� through tight junctionsinto the lumen. The resulting transepithelial osmoticgradient drives the movement of water, creating aplasma-like primary secretion.

The presence of several types of Ca2�-activated cat-ion channels in salivary acinar cells has been reported.These include a K� channel with a large single-channelconductance near 160 pS in both rodents (22) andhumans (23, 29), a 15-pS channel permeable to Na�

(22) in humans, a 30-pS Na�-permeable channel inrodents (22), a 30-pS K�-selective channel in humans,and a 40-pS K� channel in mice (14). The variety ofchannel types in different tissues and species and thelack of any specific information on the genes that codefor these channels complicate efforts to understand theroles of these various channels in fluid secretion.

As a first step toward an understanding of the role ofK� channels in salivary function, we designed an ap-proach to determine the molecular identities of theCa2�-activated K� channels in the mouse model sys-tem. Patch-clamp analysis of currents from mouse pa-rotid acinar cells revealed two levels of single-channelcurrent that were activated by increases in intracellu-lar Ca2�. On the basis of the expression profile ofchannel mRNA in parotid acinar cells, in combinationwith the electrical and pharmacological properties ofnative and recombinantly expressed channels, the na-tive, small-conductance channel is likely derived fromthe mIK1 gene. The larger conductance channel likelyreflects expression of a single “insertless” splice variantof the mSlo gene. Northern blot analysis revealed theexpression of two mSlo �-subunits in mouse parotidtissue. We compared the iberiotoxin sensitivity of thechannels in native tissue with heterologously ex-pressed channels containing each �-subunit. These re-sults and the toxin sensitivity of channels from parotidacinar cells of a �1 knockout mouse suggest that �50%of the native maxi-K channels are homotetramers of

*K. Nehrke and C. C. Quinn contributed equally to this work.Address for reprint requests and other correspondence: T. Bege-

nisich, Dept. of Pharmacology and Physiology, Box 711, Univ. ofRochester Medical Center, Rochester, NY 14642 (E-mail:ted_begenisich@urmc.rochester.edu).

The costs of publication of this article were defrayed in part by thepayment of page charges. The article must therefore be herebymarked ‘‘advertisement’’ in accordance with 18 U.S.C. Section 1734solely to indicate this fact.

Am J Physiol Cell Physiol 284: C535–C546, 2003.First published October 16, 2002; 10.1152/ajpcell.00044.2002.

0363-6143/03 $5.00 Copyright © 2003 the American Physiological Societyhttp://www.ajpcell.org C535

on July 8, 2010 ajpcell.physiology.org

Dow

nloaded from

the parotid mSlo splice variant and that the remaininglarge-conductance channels are heteromeric proteinsfrom this mSlo variant complexed with the �4-subunit.

MATERIALS AND METHODS

Western analysis. Mouse brain and parotid homogenateswere used to identify native maxi-K channel protein. Tissueswere prepared as previously described (34). After dissection,the tissues were homogenized twice by 10-s strokes at powerlevel 5 with a Polytron homogenizer (Brinkmann Instru-ments, Westbury, NY) in 5 ml of homogenization solution (10mM HEPES adjusted to pH 7.4 with Tris, 10% sucrose, 1 mMEDTA, and 1 mM PMSF) per gram of tissue, with 1 tablet ofComplete protease inhibitor (Roche Applied Science, India-napolis, IN) per 50 ml. Homogenates were centrifuged at2,500 g for 15 min at 4°C, and the supernatants were saved.The pellets were resuspended in 5 ml of homogenizationbuffer per gram of starting tissue and then homogenized andrecentrifuged as described above. The supernatants werecombined, and crude membrane proteins were precipitatedby centrifugation at 22,000 g for 20 min at 4°C. The super-natants from this step were discarded, and the pellets wereresuspended in PBS containing 1 mM EDTA, 1 mM PMSF,and Complete protease inhibitor (1 tablet per 50 ml) and thenpassed once through a 25-gauge needle and once through a30-gauge needle. Aliquots were quickly frozen in liquid N2

and stored at �85°C until use.Approximately 20 �g of crude membrane protein were

separated by two-phase Tricine polyacrylamide gel electro-phoresis (10% T/6% C resolving layer, 4% T/3% C stackinglayer) and transferred onto nitrocellulose membrane (Hy-bond ECL; Amersham Pharmacia Biotech, Piscataway, NJ)in buffer containing 10 mM 3-(cyclohexylamino)-1-propane-sulfonic acid (CAPS) adjusted to pH 11 and 10% methanol.The blot was blocked overnight at room temperature inblocking buffer (PBS containing 1% Tween 20, 4% BSA, and1% normal goat serum). After blocking, the blot was incu-bated for 2 h at room temperature in blocking buffer contain-ing a 1:250 dilution of rabbit anti-mSlo polyclonal antibody(Alomone Laboratories, Jerusalem, Israel), washed threetimes with PBS-T (PBS containing 1% Tween 20), incubatedfor 1 h at room temperature in blocking buffer containing a1:5,000 dilution of horseradish peroxidase-conjugated goatanti-rabbit IgG (Jackson ImmunoResearch Laboratories,West Grove, PA), and again washed three times with PBS-T.Immune complexes were detected on film using EnhancedChemiLuminescence (ECL; Amersham Pharmacia Biotech).

Northern blot analysis, in situ hybridization, and RT-PCR.Total RNA was prepared using Trizol reagent (Life Technol-ogies, Grand Island, NY) according to the manufacturer’sinstructions. mRNA was further purified by chromatographyon oligo(dT) resin. The RNA was fractionated by electro-phoresis in a formaldehyde-agarose gel and transferred toHybond-XL nylon membrane (Amersham). The blot was hy-bridized with a 32P-labeled double-stranded probe generatedfrom nucleotides 96–519 of the mIK1 cDNA, nucleotides1347–3591 of the mSlo cDNA, or the entire coding region ofthe �1 or �4 cDNAs in ExpressHyb solution (Clontech, PaloAlto, CA) as recommended by the manufacturer.

A subclone containing nucleotides 907–1467 of the mIK1cDNA (start codon at base 41) was used for in situ hybrid-ization analysis. In situ hybridization with the antisensecRNA probe was performed using a modification of proce-dures described by Wilkinson and Green (38). Mouse parotidgland was fixed overnight in freshly prepared ice-cold 4%paraformaldehyde in PBS. The gland was dehydrated

through ethanol into xylene and embedded in paraffin usinga Tissue-Tek V.I.P. automatic processor (Miles, Mishawaka,IN). Sections (5 �m) were adhered to commercially modifiedglass slides (Superfrost Plus; VWR, Rochester, NY), dewaxedin xylene, rehydrated through graded ethanols, and treatedwith proteinase K to enhance probe accessibility and withacetic anhydride to reduce nonspecific background. Single-stranded RNA probes were prepared by standard techniqueswith specific activities of 5 � 109 dpm/mg. Sections werehybridized at 15°C below the melting temperature, washedat high stringency (7°C below melting temperature), andtreated with RNase A to further diminish nonspecific adher-ence of the probe. Autoradiography with NBT-2 emulsion(Eastman Kodak, Rochester, NY) was performed for 25 days.Slides were developed with D19 (Eastman Kodak), and thetissue was counterstained with hematoxylin. Bright-field anddark-field images were captured with a Polaroid Digital Mi-croscope camera (Cambridge, MA) and processed usingAdobe Photoshop with Image Processing Toolkit (ReindeerGames, Asheville, NC).

For RT-PCR analysis, 1 �g of total RNA was used tosynthesize first-strand cDNA by using an RT-PCR kit (Clon-tech) according to the manufacturer’s protocol. PCR reactionswere annealed at 7°C below the melting temperature and runfor the indicated number of cycles. Oligonucleotide pairs foramplification of mSlo1, -2, and -3 (Kcnma1–3 gene products)were designed against nucleotides 1775–2066 of the mSlo1coding region, nucleotides 1–353 of the mouse expressedsequence tag (EST) BE655597, which presumably codes formSlo2 (Slack), a mouse homolog of the Caenorhabditis el-egans Slo-2 protein (39), and nucleotides 1873–2137 ofmSlo-3. Oligonucleotide pairs for amplification of the mSlo�-subunits 1 though 4 (Kcnmb1–4 gene products) were de-signed on the basis of the published sequence of the mouse �1

(nucleotides 198–426 of the open reading frame) and �4

(nucleotides 224–465 of the open reading frame) subunitsand the published sequence of the human �2 (nucleotides310–543 of the open reading frame) and �3 (nucleotides279–519 of the open reading frame) subunits. Although theproducts amplified in parentheses above code for regions thatare conserved among the �-subunits, the oligonucleotidepairs do not cross-hybridize. The efficacy of the �2 and �3

primers was assessed by using a mouse tissue control knownto be positive for those subunits (see text). All PCR bands ofan inappropriate size were cloned and sequenced and werefound to be the result of nonspecific amplification of unre-lated gene products.

Cloning strategies. Parotid mSlo splice-site variation wasanalyzed by sequencing RT-PCR products amplified frommouse parotid gland first-strand cDNA. For clarity, all nu-cleotide designations refer back to the first published mSlocoding sequence (accession no. L16912; Ref. 8). Our analysisproceeded in three steps. First, we sequenced more than 20clones from the COOH-terminal coding region containingnucleotides 1775–3591, where much of the splice-site varia-tion occurs. Next, we used 5�-RACE (rapid amplification ofcDNA ends) to determine the start codon, again sequencingmultiple clones. Finally, we amplified the full-length openreading frame and examined the sequences coding for thetransmembrane domains. This analysis suggested the pres-ence of a single parotid mSlo variant (which we have called“parSlo” to unambiguously distinguish this from the commonmSlo splice variant) that had no insertions at any of thecharacterized mSlo splice sites 1–6 (accession no. AF465344).Compared with mSlo, parSlo lacks 1) the first 81 bases,which results in the use of the “M1” alternative start codon;2) nucleotides 1979–1987 at splice site 4, located between the

C536 PAROTID K� CHANNEL GENES

AJP-Cell Physiol • VOL 284 • FEBRUARY 2003 • www.ajpcell.org

on July 8, 2010 ajpcell.physiology.org

Dow

nloaded from

eighth and ninth putative membrane-spanning segments;and 3) nucleotides 3408–3436 at splice site 6, which shiftsthe reading frame, resulting in the use of a cryptic stop codonand the formation of a “short” protein.

To clone parSlo for expression studies, we used PCR toamplify a region spanning nucleotides 776–3591 by using aNotI-tagged 3� primer and a mouse parotid cDNA template.An SphI-NotI fragment containing nucleotides 1347–3591was then digested from the PCR product and cloned into thecomplementary sites in a vector containing an mSlo variantthat is identical to parSlo from the start codon to nucleotide1347 (8). The entire coding sequence was then digested fromthe resulting vector and subcloned into pcDNA3.1 (Invitro-gen, Carlsbad, CA) using KpnI and XbaI.

The �1- and the �4-subunit coding regions were amplifiedfrom first-strand parotid cDNA by using oligonucleotidesthat included the start and stop codons and that had beentagged with a Kozak translation initiation sequence andNheI-BglII restriction sites. The resulting PCR productswere then cloned into the complementary sites of the vectorpIRES2-EGFP (Clontech). The mIK1 coding sequence wassimilarly amplified from first-strand cDNA by using taggedprimer pairs and was cloned into the NheI-BglII sites ofpIRES2-EGFP. The full coding sequences of all of the PCR-generated clones were sequenced on both strands to ensurefidelity.

Cell transfection. Chinese hamster ovary (CHO) cells andTSA cells [human embryonic kidney (HEK)-293 cells stablytransfected with the large T antigen, provided by Ronald Liand Eduardo Marban, Johns Hopkins University, Baltimore,MD] were cultured in DMEM plus 10% newborn calf serumand antibiotics in a 5% CO2 incubator at 37°C. Before trans-fection, cells were plated onto 100-mm culture dishes andgrown to �70% confluency and then transfected with a totalof 10 �g of DNA per dish by using Superfect transfectionreagent (Qiagen) following the manufacturer’s recommendedprotocol. The mIK construct was subcloned into the pIRES2-EGFP vector, which allowed identification of transfected cellswith fluorescence microscopy. The parotid mSlo variant (inpcDNA3.1) was expressed in a 1:1 ratio with the pIRES2-EGFP vector. The �1- or �4-subunits (in pIRES2-EGFP) werecotransfected with parSlo (pcDNA3.1) in a 4:1 ratio.

Electrophysiological methods. Single acinar cells were dis-sociated from mouse parotid glands with a modification of themethods previously used for rats (1). Briefly, glands weredissected from exsanguinated mice after CO2 anesthesia.Unless otherwise indicated, the data presented here are fromBlackSwiss � 129/SvJ hybrid mice. In addition, we also used�1 knockout mice (7), kindly provided by R. Brenner andR. W. Aldrich (Department of Molecular and Cellular Phys-iology, Stanford University School of Medicine, Stanford,CA), and mice (C57BL/6J) that are the wild-type “controls”for the �1 knockout animals. Glands were minced in Ca2�-free minimum essential medium (MEM; GIBCO BRL, Gaith-ersburg, MD) supplemented with 1% BSA (fraction V; SigmaChemical, St. Louis, MO). The tissue was treated for 20 min(37°C) with a 0.02% trypsin solution (MEM-Ca2� free � 1mM EDTA � 2 mM glutamine � 1% BSA). Digestion wasstopped with 2 mg/ml soybean trypsin inhibitor (SigmaChemical), and the tissue was further dispersed by twosequential 60-min treatments with collagenase (100 U/mltype CLSPA; Worthington Biochemical, Freehold, NJ) inMEM-Ca2� free plus 2 mM glutamine plus 1% BSA. Thedispersed cells were centrifuged and washed with basal me-dium Eagle (BME) (GIBCO)/BSA free. The final pellet wasresuspended in BME/BSA free plus 2 mM glutamine, and

cells were plated onto poly-L-lysine-coated glass coverslips forelectrophysiological recordings.

Whole cell and single-channel patch-clamp recordingswere done at room temperature (20–22°C) with an Axopatch1-D amplifier (Axon Instruments, Foster City, CA). Dataacquisition was performed by using a 12-bit analog-to-digitalconverter controlled by a personal computer. Whole cell K�

channel currents were measured from native parotid acinarcells and from HEK and CHO cells transfected with variouschannel clones. Except where indicated, we used an internalsolution that consisted of 135 mM K-glutamate, 10 mMHEPES (pH 7.2), and various free Ca2� concentrations es-tablished by an EGTA-Ca buffer system with 5 mM totalEGTA (Ref. 5; see also http://www.stanford.edu/�cpatton/maxc.html). The external solution consisted of (in mM) 135Na-glutamate, 5 K-glutamate, 2 CaCl2, 2 MgCl2, and 10HEPES, pH 7.2.

We measured single-channel currents from both on-celland excised, inside-out patches of parotid acinar cells. Ca2�-activated currents were stimulated in on-cell patches byapplication of 5 �M carbachol and in excised patches byapplication of the 135 K-glutamate internal solution withbuffered Ca2� levels to the inside surface of the patch. Thecell membrane voltage was nulled with an isotonic K-gluta-mate solution. The external (pipette) solution for both on-celland excised patches contained (in mM) 135 K-glutamate, 2CaCl2, 2 MgCl2, and 10 HEPES (pH 7.2). Long stretches ofdata containing openings of one to three single channels wereselected for determining amplitude histograms. These werefit with a Gaussian function to determine the single-channelcurrent levels.

We examined the dose dependence of block of current byseveral pharmacological compounds. We determined theEC50 values for these compounds by a nonlinear least-squares fit of the standard binding isotherm to the data inthe Origin software of the OriginLab (Northampton, MA).The error limits on the EC50 values are the estimated stan-dard errors of this parameter provided by the software.

Iberiotoxin (IBX) and clotrimazole were purchased fromSigma, and paxilline was from Biomol (Plymouth Meeting,PA). Charybdotoxin (CTX) was produced by expressing andpurifying a cleavable fusion protein in Escherichia coli (28,32). All solutions in experiments with IBX and CTX included30 �g/ml BSA.

RESULTS

Fluid secretion in parotid acinar cells is stimulatedby muscarinic agonists that mobilize intracellularCa2�. This process involves the activation of bothCa2�-activated Cl� and Ca2�-activated K� channels.We found significant activity of only two types of singleK� channel currents in on-cell patches of mouse pa-rotid acinar cells stimulated by 5 �M of the secreta-gogue carbachol. Examples of these currents are shownin Fig. 1, top inset. When the currents from on-cellpatches were recorded with an extracellular K� con-centration of 135 mM, the two levels were near 22 and140 pS (Fig. 1, filled squares). These same two levels ofconductance were observed from patches excised intoCa2�-containing intracellular solutions (Fig. 1, opensquares). Of eight patches, five contained both channeltypes, two contained openings only of the small-con-ductance level, and one had openings of only the large-conductance channel. The amplitude histograms (see

C537PAROTID K� CHANNEL GENES

AJP-Cell Physiol • VOL 284 • FEBRUARY 2003 • www.ajpcell.org

on July 8, 2010 ajpcell.physiology.org

Dow

nloaded from

Electrophysiological methods) showed no signs of sub-conductance states.

These conductance levels suggest that the Ca2�-acti-vated K� channels in mouse parotid acinar cells aremembers of the intermediate and maxi-K families, re-spectively. Because the intermediate-conductance K�

channels are not voltage or time dependent (15, 20, 36,37) and maxi-K channels are voltage and time dependent(13), the presence of both classes of these channels inparotid acinar cells predicts that the whole cell currentswould have two Ca2�-dependent components: one that istime and voltage independent and one that is time andvoltage dependent. Figure 2 shows that parotid acinarcells do indeed have both these components.

Figure 2A, inset, shows whole cell currents recordedfrom a parotid acinar cell patched with an internalCa2� concentration of 160 nM. Shown are currents inresponse to steps of membrane voltage to �110, �30,�10, and �50 mV from a holding potential of �70 mV.Steps to negative voltages elicited currents with littletime dependence (after the capacitative current tran-sient), whereas strong depolarizations produced cur-rents in a time- and voltage-dependent manner. Thevoltage dependence of the current measured at the endof the 40-ms pulses is shown in the main part of Fig.2A. The current-voltage relationship can be seen tocontain two components: a linear one at potentialsmore negative than about �50 mV, and an outwardrectifying component at positive potentials. The brokenline represents a linear fit to the current at the fourmost negative potentials. The interpolated zero-cur-rent potential was �78 mV, which is close to theexpected K� equilibrium potential of �82 mV for thesolutions used in this experiment. Thus the Ca2�-activated current in mouse parotid acinar cells is con-sistent with the expression of both intermediate-con-ductance and maxi-K channels.

Intermediate-conductance K� channels have a highsensitivity to the antifungal agent clotrimazole (15, 20,

36, 37). Maxi-K channels are generally resistant to thisagent (39). Thus, if these two types of channels areexpressed in parotid acinar cells, the linear, time-independent component should be inhibited by clotrim-azole with little or no effect on the nonlinear, time-dependent current. Figure 2B shows that the additionof 300 nM clotrimazole to the cell shown in Fig. 2Ainhibited the linear component by �70% with no cleareffect on the time-dependent, nonlinear current. Theaverage inhibition of the linear component by 300 nMclotrimazole was 78 � 2.8% (mean � SE, n 4).Current block in several similar experiments with arange of clotrimazole concentrations was consistentwith an EC50 value of 30 � 5.5 nM.

Thus mIK1 represents a reasonable candidate for thetime- and voltage-insensitive Ca2�-activated K� channelin parotid acinar cells. The human IK isoform, hIK1, hasbeen shown via a multiple tissue RNA dot blot to behighly expressed in salivary glands as well as in severalother nonexcitable tissues (17). Northern blot analysis ofmouse tissues with the use of a probe complementary tothe 5� portion of mIK1 detected a single transcript of �2.2kb in both colon and parotid gland, consistent with theexpected size of the mIK1 message (Fig. 3A). Subsequentin situ hybridization of semi-thin parotid gland sectionswith an antisense cRNA suggested that mIK1 expressionis limited to the acinar cells and does not occur in cells ofthe salivary ducts (Fig. 3B). Hybridization with a sensecontrol produced only a weak and nonspecific signal (datanot shown).

To solidify the identification of mIK1 as the time-and voltage-independent Ca2�-activated K� channel inparotid acinar cells, we compared several properties ofmIK1 expressed in HEK cells with the currents in thenative cells. An example of single-channel currents inan excised inside-out patch from a HEK cell trans-fected with mIK1 is shown in Fig. 1, bottom inset. Thecurrent-voltage relationship of single, expressed mIK1channels (Fig. 1, open triangles) was indistinguishable

Fig. 1. Single-channel currents from native parotidacinar cells and from cells expressing mIK1 andparSlo. Single-channel current-voltage relationshipsare shown for large- and small-conductance channelsrecorded from on-cell patches of parotid acinar cells(■ ); excised, inside-out patches from acinar cells (�);inside-out patches from HEK cells transfected withmIK1 (‚); inside-out patches from human embryonickidney (HEK) cells transfected with parSlo (E); andinside-out patches from Chinese hamster ovary(CHO) cells transfected with parSlo (ƒ). For the on-cell patches, channels were activated with 5 �M car-bachol. Channels in excised patches were activated by160, 300, or 400 nM Ca2�. Inset: examples of thesmall- and large-conductance channels observed inon-cell patches on parotid acinar cells recorded at �70mV (top) and examples of currents from inside-outpatches from HEK cells transfected with mIK1 andparSlo channels and recorded at �80 and �75 mV,respectively (bottom). Calibration: 5 pA for all recordsand 20 ms for parotid channels and mSlo, 100 ms formIK1.

C538 PAROTID K� CHANNEL GENES

AJP-Cell Physiol • VOL 284 • FEBRUARY 2003 • www.ajpcell.org

on July 8, 2010 ajpcell.physiology.org

Dow

nloaded from

from the native, small-conductance channels. Figure 2C,inset, shows that expressed mIK1 had little voltage andtime dependence, and the main part of Fig. 2C shows thatexpressed mIK1 channels were quite sensitive to 300 nMclotrimazole. Several similar experiments with a range ofclotrimazole concentrations were consistent with inhibi-tion of recombinantly expressed mIK1 channel by thiscompound with an EC50 of 18 � 5 nM, a value a bitsmaller than that for the linear component of native cells.

We also compared the tetraethylammonium (TEA)sensitivity of the linear component of native cell cur-rent to that of heterologously expressed mIK1 chan-nels. The native current was inhibited by TEA with anEC50 value of 6 � 1 mM. Current through expressedmIK1 channels was inhibited with a similar EC50

value of 8.7 � 1 mM (data not shown).Thus we found mIK1 to be robustly expressed in

mouse parotid acinar cells, and many properties ofexpressed mIK1 channels were quite similar to those ofthe linear, Ca2�-activated K� current in native cells.We used a similar approach to test the possibility thata product of the Slo gene underlies the time- andvoltage-dependent Ca2�-activated current in mouseparotid acinar cells.

Fig. 2. Comparison of time- and voltage-independent K� current in native cells with expressed mIK1. A, inset: rawcurrents from a parotid acinar cell in response to step changes in the membrane voltage from a �70-mV holdingpotential to the indicated values. Main graph in A shows currents measured at the end of the test pulses (■ ) to theindicated membrane potentials (Vm). Dotted line represents linear least-squares fit to the 4 most negative voltages.B: data from the same cell as in A in the presence of 300 nM clotrimazole. C, inset: raw currents from a HEK celltransfected with mIK1. Main graph in C shows current-voltage relationship of expressed mIK1 current before (■ ),during (F), and after washout (�) of 300 nM clotrimazole.

Fig. 3. Northern blot and in situ analysis of parotid mRNA. A:Northern analysis of RNA isolated from parotid gland and from colon(as a control) reveals a single transcript of �2.2 kb. B: in situanalysis. Parotid glands were fixed and sectioned at 5 �M and thenprobed with an anti-sense mIK1 cRNA. Labeling (red) occurs in theacinar but not ductal cells (outlined in black at �20 magnification).

C539PAROTID K� CHANNEL GENES

AJP-Cell Physiol • VOL 284 • FEBRUARY 2003 • www.ajpcell.org

on July 8, 2010 ajpcell.physiology.org

Dow

nloaded from

The Slo1 channel was originally cloned from Dro-sophila (2) and subsequently from mouse and humansources (8, 33). Alternative splicing of the Slo1 geneproduct in mice (Kcnma1 gene) and humans (KCNMA1gene) results in multiple variants with distinct expres-sion patterns and calcium sensitivities (4, 8, 18, 31).Furthermore, two additional Slo isoforms are encodedby separate genes. Slo2 was first isolated from C.elegans and is sensitive to chloride as well as calcium(40), whereas Slo3 expression is restricted to the testesin mice and humans and is insensitive to calcium butsensitive to intracellular pH (30). Both Slo2 and Slo3homologs can be found in the human genome databaseof predicted coding regions and in the mouse ESTdatabase. RT-PCR analysis of Slo isoforms in mouseparotid gland revealed that expression was limited tothe mSlo1 isoform; the amplification of products of thepredicted size from kidney and testes cDNA confirmedthe efficacy of the mSlo 2 and 3 primer pairs, respec-tively (Fig. 4). The presence of mSlo was tested fordirectly by performing Northern and Western blotanalyses of parotid gland mRNA and membrane pro-teins, respectively (Fig. 4, B and C). A single transcript

of �5.3 kb was recognized by an mSlo-specific cDNAprobe (Fig. 4B), whereas a commercially available anti-mSlo antibody reacted with a protein with an apparentmolecular mass of 115 kDa in both brain and parotidmembrane preparations (Fig. 4C). The brain mSlo pro-tein appeared to be slightly larger than the parotidversion, consistent with the splice variation that weobserved in the parotid gland (see below). Preincuba-tion of the antibody with a purified peptide consistingof the mSlo recognition site resulted in a loss of mSlo-specific signal. Finally, in situ analysis of parotid sec-tions failed to detect mSlo (data not shown), suggestingthat the message may be expressed at levels below thedetection sensitivity of this method.

Alternative splicing of the Slo channel transcriptresults in a molecular diversity that can alter thevoltage and calcium sensitivity of the protein in manyspecies (4, 8, 18, 26, 31). To date, at least six splice siteshave been characterized in the coding region of themouse transcript, and the beginning of the transcript,

Fig. 4. Expression of mSlo-1, -2, and -3 (Kcnma1–3 genes) in theparotid gland. A: RT-PCR was used to amplify mouse parotid glandcDNA with oligonucleotide pairs specific for the mouse large-conduc-tance Ca2�-activated K� channel isoforms 1, 2, and 3 for the indi-cated number of cycles. mSlo1 was expressed in the parotid gland,whereas mSlo2 and 3 were not. Control reactions indicated thatisoforms 2 and 3 were expressed in the kidney and testes, respec-tively. B: Northern blot of parotid mRNA hybridized to a probe formSlo. The probe detected a band of �5.3 kb, as expected. C: Westernblot of brain (lane 1) and parotid (lane 2) membrane proteins withanti-mSlo antibody (Alamone Laboratories). A band of �115 kDawas detected in both preparations, though the protein from brainappeared to be slightly larger, as predicted from cDNA splice siteanalysis. Preincubation with a peptide containing the antibody rec-ognition sequence abolished the signal (lane 3).

Fig. 5. Splice sites in the mSlo gene. A: schematic of the mSlo codingregion and splice sites. Putative transmembrane domains S1–S10are indicated by boxes, characterized splice sites 1–6 are indicatedby arrows, and the pore region is denoted as P. The NH2 terminus ofthe protein (‡) is predicted to be highly diverse based on 5� RACE (8).The splice sites that were found to be altered in the parSlo variantare indicated by an asterick B: sequence alignment of hSlo (GenBankaccession no. NM_002247), mSlo (accession no. L16912), and parSlo(accession no. AF465244), the parotid mSlo splice variant. Portionsof the proteins that are identical were left out of the alignment andare shown only as dotted lines. The regions where these proteinsdiffer are indicated in bold, with the particular splice sites respon-sible for the changes noted. Single site differences between hSlo andparSlo are also highlighted by an asterick. The major differencesbetween mSlo and parSlo consisted of a shorter amino terminus, theabsence of the IYF motif at splice site 3, and a retained exon at splicesite 6 that contains a premature stop site.

C540 PAROTID K� CHANNEL GENES

AJP-Cell Physiol • VOL 284 • FEBRUARY 2003 • www.ajpcell.org

on July 8, 2010 ajpcell.physiology.org

Dow

nloaded from

including the start of the open reading frame, has beenshown to be highly diverse (Ref. 8; Fig. 5A). To assessthe distribution of mSlo1 splice variants in the parotidgland, the coding region for the COOH terminus of theprotein, where most of the identified splice sites occur,was amplified using RT-PCR and multiple clones weresequenced. All of the clones were found to be identicaland insertless at each of the characterized splice sites.Additional 5�-RACE and RT-PCR of the full-lengthclone suggested that a single mSlo1 transcript was ex-pressed in the parotid gland (data not shown). This tran-script corresponded to a short version of the protein,resulting from an in-frame stop codon following the finalsplice site. This splice variant is highly homologous to thehuman Slo1 mRNA (accession no. NM_002247) and dif-fers primarily in the last eight conceptually translatedamino acids, as shown in Fig. 5B.

We compared several properties of the mouse parotidmSlo variant (parSlo) expressed in CHO and HEK cellswith the time- and voltage-dependent current in nativeparotid cells. Figure 1, bottom inset, contains an exam-ple of single-channel currents in an inside-out patchfrom a HEK cell transfected with the parSlo. The single-channel current-voltage relationship of parSlo channelsexpressed in HEK (Fig. 1, open circles) and CHO cells(Fig. 1, open inverted triangles) was similar to that of thelarge-conductance channels in the native cells.

Maxi-K channels are sensitive to the scorpion toxinsCTX and IBX (13). Expressed IK1 channels are alsosensitive to CTX but are relatively insensitive to IBX(15, 36). Thus IBX is a specific pharmacological tool formaxi-K channels, so we examined IBX block of native

currents and recombinantly-expressed mSlo channels.Figure 6A, inset, contains an example of the time- andvoltage-dependent currents of native parotid acinarcells recorded in the presence of 300 nM clotrimazole toinhibit the linear mIK1 currents. The main part of Fig.6A shows that application of 500 nM IBX had a modesteffect on the current recorded at �50 mV. In thisexample, 500 nM IBX blocked �33% of the current.The average amount of block of native current by thisconcentration was 38 � 2.1% (n 6).

Heterologously expressed parSlo produced time- andvoltage-dependent currents (Fig. 6B, inset) that weregenerally similar to the currents in native cells. How-ever, though not investigated in detail, the time courseof activation of expressed parSlo was somewhat fasterthan the current in the native cells. Significantly, ex-pressed parSlo channels were considerably more sen-sitive to IBX than the native current. The main part ofFig. 6B shows that 100 nM IBX blocked 95% of thecurrent (at �50 mV) through parSlo channels heter-ologously expressed in CHO cells. Similar data with2–100 nM IBX were fit by a standard binding isothermwith an EC50 value of 6.1 � 2 nM for IBX block ofhomomeric parSlo channels expressed in HEK cells.

The good agreement between the single-channel con-ductance coupled with the discrepancy between theIBX sensitivity of the native current and expressedparSlo raises the possibility that the native channelsmay contain an auxiliary subunit in addition to theproduct of the parSlo gene. Four �-subunits that canassemble with Slo proteins have been identified (3, 6, 9,10, 24, 35), and the presence of these �-subunits alters

Fig. 6. Iberiotoxin block of parotid andexpressed K� channels. A: native pa-rotid acinar cells treated with 300 nMclotrimazole to minimize mIK1 chan-nel activity. Current was activated by160 nM intracellular Ca2�. Currentwas measured at the end of pulses to�50 mV. Application of 500 nM iberio-toxin (IBX) is indicated by the arrow.B: data from a cell transfected withparSlo only. Application of 100 nM IBXis indicated by the arrow. C: data froma cell cotransfected with parSlo��1.Applications of 500 nM IBX and 5 mMTEA are indicated by arrows. D: datafrom an HEK cell cotransfected withparSlo��4. Applications of 500 and2,500 nM IBX and 5 mM TEA are in-dicated by arrows. All currents (exceptthose in A) are from CHO cells patchedwith 400 nM Ca2�. Insets: raw currentsin response to steps of membrane volt-age to �10, �30, and �50 mV. Calibra-tions: 2 nA, 10 ms (A); 0.5 nA, 10 ms(B); 1 nA, 10 ms (C); 0.25 nA, 20 ms (D).

C541PAROTID K� CHANNEL GENES

AJP-Cell Physiol • VOL 284 • FEBRUARY 2003 • www.ajpcell.org

on July 8, 2010 ajpcell.physiology.org

Dow

nloaded from

the scorpion toxin sensitivity of the heteromeric chan-nels (3, 10, 24).

To test for the presence of these �-subunits in themouse parotid gland, we designed primer pairs forRT-PCR analysis based on multiple sequence align-ments of the published �-subunit gene products (hu-man KCNMB1–4; mouse Kcnmb1 and 4). The followinggoals applied: first, to flank homologous segments ofthe �-subunit coding regions; second, to prevent cross-hybridization; and third, to end the primers on codonsthat represent highly conserved amino acids. Thisstrategy was expected to yield primer pairs for �2- and�3-subunits that would recognize both the human andmouse homologs, because the mouse cDNA sequenceshad yet to be reported. Our results indicated that �1-and �4-subunits were most prominently expressed inmouse parotid gland, whereas the �2 and �3 amplimersets could be validated for use in mice by using mousekidney and trachea cDNA templates, respectively (Fig.7). Furthermore, successive rounds of amplificationfailed to result in the accumulation of specific PCRproducts for �2- or �3-subunits from parotid cDNA,suggesting that their expression levels are below thedetection limits of our assay (data not shown).

Expression of the �1- and �4-subunit mRNA in pa-rotid gland was demonstrated directly by Northernblot analysis (Fig. 7B). We used a commercially avail-able anti-�1 antibody but were unable to detect a spe-cific signal in a preparation of parotid gland mem-branes by Western blot analysis (data not shown).Similarly, attempts to localize the signal to specificcells in the parotid gland through immunohistochem-istry were unsuccessful. Thus molecular studies alonecannot determine which of the two �-subunits, if ei-ther, are expressed in parotid acinar cells. To answerthis question, we compared the electrophysiologicaland pharmacological characteristics of the parSlochannel coexpressed with either �1 or �4 to the endog-enous parotid acinar cell current.

Heteromeric parSlo��1 channels expressed voltage-and time-dependent currents (Fig. 6C, inset) generallysimilar to the currents in native cells. These hetero-meric channels had a single-channel conductance in-distinguishable from that of homomeric parSlo chan-nels (not shown) consistent with results with hSlo andthe human �1-subunit (27). However, the addition ofthe �1-subunit substantially reduced the IBX sensitiv-ity of the channels, as shown in the main part of Fig.6C. Even 500 nM IBX had only a modest effect onheteromeric parSlo��1: an average block at �50 mV of17.5 � 7% (n 5). This degree of block is consistentwith an EC50 value of 2,400 nM, and in agreement withthis value, 2,500 nM blocked 54 and 47% of the current intwo independent experiments. Heteromeric parSlo��1channels remained sensitive to external TEA, as shownin Fig. 6C.

Coexpression of parSlo and �4 also produced time-and voltage-dependent currents resembling those inthe native cells (Fig. 6D, inset) with the same single-channel conductance as the native channels and homo-meric parSlo channels expressed in HEK and CHOcells (data not shown). As shown in the main part ofFig. 6D, heteromeric parSlo��4 channels were entirelyinsensitive to 500 nM IBX. This concentration of toxinreduced the current (at �50 mV) through the hetero-meric channels expressed in CHO cells by 1.9 � 1.6%(n 3). Application of 2,500 nM for 10 min (Fig. 6D)had no more effect than 500 nM (P 0.1, n 3). Asshown, these heteromeric parSlo��4 channels re-mained sensitive to external TEA.

Thus the voltage- and time-dependent current inmouse parotid acinar cells did not have the IBX sensi-tivity of homomeric parSlo channels or that of hetero-meric parSlo��1 or �4 channels, raising the possibilitythat the native current could include two populationsof channels: one that is relatively sensitive to the toxin(like homomeric parSlo channels), and one that is toxinresistant (like heteromeric parSlo��1 or �4 channels).The data in Fig. 8A show that only �39% of the nativecurrent (in BlackSwiss � 129/SvJ mice) was sensitiveto IBX. This IBX-sensitive component had an EC50value of �31 � 14 nM. About 61% of the voltage- andtime-dependent current in mouse parotid acinar cellswas insensitive to IBX up to concentrations as large as2,500 nM. We also examined the CTX sensitivity of the

Fig. 7. Expression of mSlo �-subunits 1–4 (Kcnmb1–4 genes) in theparotid gland. A: RT-PCR was used to amplify mouse parotid glandcDNA with oligonucleotide pairs specific for the mouse large-conduc-tance Ca2�-activated K� channel �-subunit isoforms 1–4 for theindicated number of cycles. �1 and �4 were expressed in the mouseparotid gland, whereas �2 and �3 were not. Control reactions indi-cated that �2 was expressed in the mouse kidney and �3 was ex-pressed in the trachea. B: Northern blot of parotid mRNA hybridizedto a probe for maxi-K �1. The probe detected a single band of �1.7 kb.C: Northern blot of parotid mRNA hybridized to a probe for maxi-K�4. This probe hybridized with an mRNA of �1.45 kb.

C542 PAROTID K� CHANNEL GENES

AJP-Cell Physiol • VOL 284 • FEBRUARY 2003 • www.ajpcell.org

on July 8, 2010 ajpcell.physiology.org

Dow

nloaded from

channels in native cells and found that 55 � 2% of thecurrent was sensitive to this toxin with an EC50 valueof 13 � 2 nM (data not shown); �45% of the currentwas insensitive to CTX up to levels as high as 500 nM.

Even though only 40–45% of the native time-, volt-age-, and Ca2�-dependent current was sensitive toeven very high IBX and CTX concentrations, much, ifnot all, of the current could be blocked by externalTEA, as shown in Fig. 6, C and D. In addition, essen-tially all the current could be inhibited by paxilline.This diterpene mycotoxin is a very potent nonpetidylinhibitor of maxi-K channels (reviewed in Ref. 13). Inthe example in Fig. 8B, 1 �M paxilline blocked 96% ofthe time- and voltage-dependent current in the pres-ence of 160 nM intracellular Ca2� (and 300 nM clotrim-azole to inhibit current through mIK1 channels). Theaverage inhibition in five such experiments was 94 �1.9%. This result and those with external TEA indicatethat 100% of the current recorded with these condi-tions is through maxi-K channels even though not allthe current can be blocked by even very high concen-trations of the scorpion toxins.

The moderate sensitivity of heteromeric parSlo��1channels to IBX (see Fig. 6C and related text above)suggests that these channels do not contribute signif-icantly to the total maxi-K current in native parotidacinar cells. However, the slow onset and small block ofthe heterologously expressed heteromeric channels di-minishes the strength of any such conclusion. An al-ternate method to test the idea that heteromericparSlo��1 channels do not contribute significantly tothe maxi-K current in native cells is to determine theIBX sensitivity of the current in cells from mice inwhich the �1 gene has been disrupted (7). The results ofsuch a test are presented in Fig. 9.

Figure 9 shows the amount of IBX block of maxi-Kchannels in �1 knockout mice (filled squares) alongwith similar data from their wild-type counterparts(open circles). There appears to be little differencebetween the IBX block of wild-type and �1 knockout

Fig. 8. IBX and paxilline block of maxi-K chan-nels in mouse parotid acinar cells. A: concentra-tion dependence of the amount of block of parotidacinar cell currents by the indicated concentra-tions of IBX. Ca2�-activated current induced by160 nM Ca2� (except for 1 experiment with 250nM). The time- and voltage-insensitive currentwas minimized by application of 300 nM clotrim-azole. Values are means � SE from 3–6 mea-surements. Line represents fit of a standardbinding isotherm with EC50 and maximum blockvalues of 31 � 14 nM and 39 � 3.2%, respec-tively. B: paxilline block of maxi-K channels inmouse parotid acinar cells. Currents were mea-sured at the end of 40-ms pulses to the voltagesindicated from a �70-mV holding potential inthe absence (control) and presence of 1 �M pax-illine. Inset: raw currents from 40-ms steps tomembrane potentials of �110, �30, 10, and 50mV from a holding voltage of �70 mV in theabsence (control) and presence of 1 �M paxillineas indicated.

Fig. 9. IBX block of maxi-K channels from �1 knockout mice. Con-centration dependence of the amount of block of parotid acinar cellcurrents by the indicated concentrations of IBX is shown. Ca2�-activated current induced by 160 nM Ca2�. The time- and voltage-insensitive current was minimized by application of 300 nM clotrim-azole. Values are means � SE from 3–4 measurements. Data arefrom C57BL/6J (E) and �1 knockout animals (■ ). Line represents fitof a standard binding isotherm with EC50 and maximum blockvalues of 22 � 6 nM and 56 � 2%, respectively.

C543PAROTID K� CHANNEL GENES

AJP-Cell Physiol • VOL 284 • FEBRUARY 2003 • www.ajpcell.org

on July 8, 2010 ajpcell.physiology.org

Dow

nloaded from

animals, and all the data are consistent with 56 � 2%of the current sensitive to IBX with an EC50 value of22 � 6 nM; that is, there is �50% of the maxi-K currentin mouse parotid that is insensitive to IBX whether ornot the �1 gene is active, confirming the suggestionthat heteromeric parSlo��1 channels do not signifi-cantly contribute to the maxi-K current of mouse pa-rotid acinar cells.

DISCUSSION

There have been multiple studies of K� channels inseveral types of salivary glands from different species(12, 14, 16, 22, 29). The results of these various studiessuggest that the K� channel expression pattern may bespecific to particular glands and particular species.Therefore, a complete understanding of the physiolog-ical role of K� channels in salivary glands requires asystematic approach that includes identifying the rel-evant genes in a single model system as well as aquantitative comparison of the electrophysiologicalproperties of the native channels with those of ex-pressed, candidate genes.

Developments in targeted gene ablation technologyhave made the mouse an attractive system in which toexamine the physiological role of ion transport proteinsin transmembrane fluid dynamics (25). We found thatmouse parotid acinar cells express significant amountof only two types of Ca2�-activated K� channels: 1) arelatively time- and voltage-independent K-selectivechannel that has a single-channel conductance (in 135mM external K�) of �22 pS, and 2) a voltage- andtime-dependent K� channel with a single-channel con-ductance of �140 pS. These properties suggestedKcnn4 (mIK1) and Kcnma (mSlo) as candidate genescoding for these channels. Thus we designed experi-ments to determine whether these genes were ex-pressed in mouse parotid acinar cells and comparedseveral properties of the native channels with those ofthe heterologously expressed candidate genes.

Northern blot and in situ analyses showed thatmIK1 was abundantly expressed in acinar but notductal cells of the mouse parotid gland, consistent witha primary role in secretion in this tissue. This finding isconsistent with the observation that salivary glandsexpress the highest detectable levels of mIK1 tran-script among all of the organs surveyed (16). Othersources of mIK1 include nonexcitable tissues such aslung, trachea, kidney, and stomach. The distribution ofmIK1 in tissues rich in epithelia supports a significantrole for mIK1 in secretion or electrolyte reabsorption.

Heterologously expressed mIK1 channels have sev-eral properties in common with the voltage- and time-independent K� current in mouse parotid acinar cells:1) a single-channel conductance indistinguishable fromthat of the native current, 2) a similar lack of voltageand time dependence, 3) a similar sensitivity to clo-trimazole, and 4) a nearly identical sensitivity to TEA.

Because the voltage- and time-dependent componentof K� current in the native cells appeared to have theproperties of a maxi-K channel, we tested for the ex-

pression of genes known to code for these types ofchannels. The gene family to which Slo1, the originalmaxi-K channel, belongs is composed of two additionalmembers, Slo2 and Slo3. As expected on the basis oftheir published distribution (mSlo3 is testes specific)and activities (mSlo2 is chloride dependent; mSlo3 ispH sensitive/Ca2� insensitive), the results of RT-PCRanalysis suggest that mSlo1 is the most prominentlyexpressed member of this gene family in the parotidgland.

The mSlo1 gene has been shown to have alterna-tively spliced products, and many of the splice variantsexhibit distinct electrophysiological characteristics andgreatly differing Ca2� sensitivities (4, 8, 18, 31). Weassayed the most common splice sites in the NH2 andCOOH termini of the protein and found that a singlehomogenous insertless variant was present in the pa-rotid gland. Although this variant has previously beendescribed in mice (31), it has not been assayed electro-physiologically. Instead, a separate variant, mbr5 (8),was expressed for patch-clamp analysis under the as-sumption that it was insertless (31). However, thembr5 variant contains a three-amino acid insertion(IYF) at the second splice site in the COOH terminusand, in addition, codes for the “long” form of the pro-tein. The true insertless form of mSlo ends prema-turely because of a cryptic stop codon brought intoframe by the splicing of nucleotides 3408–3436 fromthe mSlo coding region. Interestingly, the insertlessmSlo variant found in the parotid most closely resem-bles one of the published human Slo sequence variants(11, 27, 33).

The mouse parotid mSlo variant (parSlo) expressed inHEK and CHO cells had a single-channel conductancethat was indistinguishable from the large-conductancechannel in the native cells. However, heterologously ex-pressed parSlo channels were substantially more sensi-tive to IBX than was the current in the native cells. Thuswe tested for the presence of four mSlo �-subunits knownto alter scorpion toxin sensitivity. We found that themouse parotid expressed only the �1- and �4-subunits,and we examined the IBX sensitivity of parSlo coex-pressed with each of these subunits.

We found that both the mouse �1- and �4-subunitsreduce the IBX sensitivity of the mouse parotid mSlovariant parSlo, with a larger effect produced by the�4-subunit: heteromeric parSlo��4 channels were in-sensitive to IBX concentrations as large as 2,500 nM.This is the first examination of the actions of IBX onthis particular mSlo splice variant in heteromericchannels with the mouse �-subunits, so no direct com-parison to existing data is possible. However, our re-sults are qualitatively similar to those obtained withhuman Slo (hSlo) channels coexpressed with human�1- and �4-subunits (3, 10, 24), although the coassem-bly of the mouse parotid mSlo variant with the �-sub-units appears to more substantially decrease IBX sen-sitivity.

We found that the IBX and CTX dose-response rela-tionships of the native channels (Figs. 8A and 9) wereconsistent with two populations of maxi-K type chan-

C544 PAROTID K� CHANNEL GENES

AJP-Cell Physiol • VOL 284 • FEBRUARY 2003 • www.ajpcell.org

on July 8, 2010 ajpcell.physiology.org

Dow

nloaded from

nels in the native cells: one (�50% of the total) with areasonably high affinity for toxin, and another popula-tion that is essentially insensitive to the toxin (to 500nM CTX and 2,500 nM IBX). Because we found hetero-meric parSlo��1 channels to be moderately sensitiveto IBX and hSlo��1 are quite CTX sensitive (3, 24), itseems reasonable to consider that the maxi-K, Ca2�-activated K� current in native parotid is composed ofapproximately equal numbers of homotetramericparSlo channels and heteromeric parSlo��4 channels.This suggestion is strengthened by our results showingthat �50% of the maxi-K current in parotid acinar cellsfrom �1 knockout mice remained insensitive to 2,500nM IBX. Although the �1 gene ablation results could becompromised by an exact compensation by �4 expres-sion, the most parsimonious explanation that accountsfor all the data is that mouse parotid acinar cellsexpress equal numbers of homotetrameric parSlo chan-nels and heteromeric parSlo��4 channels.

We thank Keerang Park for cloning the mIK1 from mouse parotidtissue and Paul D. Kingsley for assistance with the in situ analysis.We are grateful to J. E. Melvin for considerable discussions of thiswork and for critically reading the manuscript. We thank J. Thomp-son for assistance with the electrophysiological experiments and forcritically reading the manuscript and Jodi Pilato and PamelaMcPherson for the preparation of parotid acinar cells.

This work was supported by National Institute of Dental ResearchGrants DE-13539 (T. Begenisich) and DE-14119 (K. Nehrke).

REFERENCES

1. Arreola J, Melvin JE, and Begenisich T. Activation of cal-cium-dependent chloride channels in rat parotid acinar cells.J Gen Physiol 108: 35–47, 1996.

2. Atkinson NS, Robertson GA, and Ganetzky B. A componentof calcium-activated potassium channels encoded by the Dro-sophila slo locus. Science 253: 551–555, 1991.

3. Behrens R, Nolting A, Reimann F, Schwarz M, WaldschutzR, and Pongs O. hKCNMB3 and hKCNMB4, cloning and char-acterization of two members of the large-conductance calcium-activated potassium channel � subunit family. FEBS Lett 474:99–106, 2000.

4. Benkusky NA, Fergus DJ, Zucchero TM, and England SK.Regulation of the Ca2�-sensitive domains of the maxi-K channelin the mouse myometrium during gestation. J Biol Chem 275:27712–27719, 2000.

5. Bers D, Patton C, and Nuccitelli R. A practical guide to thepreparation of Ca buffers. Methods Cell Biol 40:3–29, 1994.

6. Brenner R, Jegla TJ, Wickenden A, Liu Y, and Aldrich RW.Cloning and functional characterization of novel large conduc-tance calcium-activated potassium channel � subunits, hKC-NMB3 and hKCNMB4. J Biol Chem 275: 6453–6461, 2000.

7. Brenner R, Perez GJ, Bonev AD, Eckman D, Kosek JC,Willer SW, Patterson AJ, Nelson MT, and Aldrich RW.Vasoregulation by the �1 subunit of the calcium-activated potas-sium channel. Nature 407:870–876, 2000.

8. Butler A, Tsunoda S, McCobb DP, Wei A, and Salkoff L.mSlo, a complex mouse gene encoding “maxi” calcium-activatedpotassium channels. Science 261: 221–224, 1993.

9. Chang CP, Dworetsky SI, Wang J, and Goldstein ME.Differential expression of the � and � subunits of the large-conductance calcium-activated potassium channel: implicationsfor channel diversity. Mol Brain Res 45: 33–40, 1997.

10. Dworetsky SI, Boissard CG, Lum-Ragan JT, McKay MC,Post-Munson DJ, Trojnacki JT, Chang CP, and GribkoffVK. Phenotypic alteraction of a human BK (hSlo) channel byhSlo� subunit coexpression: changes in blocker sensitivity, acti-vation/relaxation and inactivation kinetics, and protein kinase Amodulation. J Neurosci 16:4543–4550, 1996.

11. Dworetzky SI, Trojnacki JT, and Gribkoff VK. Cloning andexpression of a human large-conductance calcium-activated po-tassium channel. Mol Brain Res 27: 189–193, 1994.

12. Gallacher DV and Morris AP. A patch-clamp study of potas-sium currents in resting acetylcholine-stimulated mouse sub-mandibular acinar cells. J Physiol 373: 379–395, 1986.

13. Gribkoff VK, Starrett JE Jr, and Dworetzky SI. The phar-macology and molecular biology of large-conductance calcium-activated (BK) potassium channels. Adv Pharmacol 17: 319–348, 1997.

14. Hayashi T, Young JA, and Cook DI. The Ach-evoked Ca2�-activated K� current in mouse mandibular secretory cells. Sin-gle channel studies. J Membr Biol 151: 19–27, 1996.

15. Ishii TM, Silvia C, Hirschberg B, Bond CT, Adelman JP,and Maylie J. A human intermediate conductance calcium-activated potassium channel. Proc Natl Acad Sci USA 94:11651–11656, 1997.

16. Ishikawa T and Murakami M. Tetraethylammonium-insensi-tive, Ca2�-activated whole-cell K� currents in rat submandibu-lar acinar cells. Pflugers Arch 428: 516–525, 1995.

17. Jensen BS, Strobaek D, Christophersen P, Jorgensen TD,Hansen C, Silahtaroglu A, Olesen SP, and Ahring PK.Characterization of the cloned human intermediate-conductanceCa2�-activated K� channel. Am J Physiol Cell Physiol 275:C848–C856, 1998.

18. Lagrutta A, Shen KZ, North RA, and Adelman JP. Func-tional differences among alternatively spliced variants of Slow-poke, a Drosophila calcium-activated potassium channel. J BiolChem 269: 20347–20351, 1994.

19. Lau KR, Howorth AJ, and Case RM. Identification and reg-ulation of K� and Cl� channels in human parotid acinar cells.J Physiol 425: 407–427, 1990.

20. Logsdon NJ, Kang J, Togo JA, Christian EP, and Aiyar J.A novel gene, hKCa4, encodes the calcium-activated potassiumchannel in human T lymphocytes. J Biol Chem 272: 32723–32726, 1997.

21. Martinez JR, Cassity N, and Reed P. An examination offunctional linkage between K efflux and 36Cl efflux in rat sub-mandibular gland in vitro. Arch Oral Biol 32: 891–895, 1987.

22. Maruyama Y, Gallacher DV, and Petersen OH. Voltage andCa2�-activated K� channel in basolateral acinar cell membranesin mammalian salivary glands. Nature 302: 827–829, 1983.

23. Maruyama Y, Nishiyama A, and Teshima T. Two types ofcation channels in the basolateral cell membrane of humansalivary gland acinar cells. Jpn J Physiol 36: 219–223, 1986.

24. Meera P, Wallner M, and Toro L. A neuronal � subunit(KCNMB4) makes the large conductance, voltage- and Ca2�-activated K� channel resistant to charybdotoxin and iberiotoxin.Proc Natl Acad Sci USA 97: 5562–5567, 2000.

25. Melvin JE, Nguyen HV, Evans RL, and Shull GE. What cantransgenic and gene targeted mouse models teach us aboutsalivary gland physiology? Adv Dent Res 14: 5–11, 2000.

26. Navaratnam DS, Bell TJ, Tu TD, Cohen EL, and Ober-holtzer JC. Differential distribution of Ca2�-activated K� chan-nel splice variants among hair cells along the tonotopic axis ofthe chick cochlea. Neuron 19: 1077–1085, 1997.

27. Pallanck L and Ganetzky B. Cloning and characterization ofhuman and mouse homologs of the Drosophila calcium-activatedpotassium channel gene, slowpoke. Hum Mol Genet 3: 1239–1243, 1994.

28. Park CS, Hausdorff SF, and Miller C. Design, synthesis, andfunctional expression of a gene for charybdotoxin, a peptideblocker of K� channels. Proc Natl Acad Sci USA 88: 2046–2050,1991.

29. Park K, Case RM, and Brown PD. Identification and regula-tion of K� and Cl� channels in human parotid acinar cells. ArchOral Biol 46: 801–810, 2001.

30. Schreiber M, Wei A, Yuan A, Gaut J, Saito M, and SalkoffL. Slo3, a novel pH-sensitive K� channel from mammalianspermatocytes. J Biol Chem 273: 3509–3516, 1998.

31. Shipston MJ, Duncan RR, Clark AG, Antoni FA, and TianL. Molecular components of large conductance calcium-activatedpotassium (BK) channels in mouse pituitary corticotropes. MolEndocrinol 13: 1728–1737, 1999.

C545PAROTID K� CHANNEL GENES

AJP-Cell Physiol • VOL 284 • FEBRUARY 2003 • www.ajpcell.org

on July 8, 2010 ajpcell.physiology.org

Dow

nloaded from

32. Stampe P, Kolmakova-Partensky L, and Miller C. Intimationsof K� channel structure from a complete functional map of themolecular surface of charybdotoxin. Biochemistry 33:443–450.

33. Tseng-Crank J, Foster CD, Krause JD, Mertz R, GodinotN, DiChiara TJ, and Reinhart PH. Cloning, expression, anddistribution of functionally distinct Ca2�-activated K� channelisoforms from human brain. Neuron 13: 1315–1330, 1994.

34. Turner RJ, George JN, and Baum BJ. Evidence for a Na�/K�/Cl� cotransport system in basolateral membrane vesiclesfrom the rabbit parotid. J Membr Biol 94: 143–152, 1986.

35. Uebele VN, Lagrutta A, Wade T, Figueroa DJ, Liu Y, Mc-Kenna E, Austin CP, Bennett PB, and Swanson R. Cloningand functional expression of two families of �-subunits of thelarge-conductance calcium-activated K� channel. J Biol Chem275: 23211–23218, 2000.

36. Vandorpe DH, Shmukler BE, Jiang L, Lim B, Maylie J,Adelman JP, de Franceschi L, Cappellini MD, BrugnaraC, and Alper SL. cDNA cloning and functional expression of the

mouse Ca2�-gated channel, mIK1. J Biol Chem 273: 21542–21553, 1998.

37. Warth R, Hamm K, Bleich M, Kunzelmann K, von Hahn T,Screiber R, Ullrich E, Mengel M, Trautmann N, Kindle P,Schwab A, and Greger R. Molecular and functional character-ization of the Ca2�-regulated K� channel (rSK4) of coloniccrypts. Pflugers Arch 438: 437–444, 1999.

38. Wilkinson DG and Green J. Postimplantation MammalianEmbryos: A Practical Approach, edited by Copp AJ and CockroftDL. New York: Oxford University Press, 1990, p. 155–171.

39. Wulff H, Miller MJ, Hansel W, Grissmer S, Cahalan MD,and Chandy KG. Design of a potent and selective inhibitor ofthe intermediate-conductance Ca2�-activated K� channel,IKCa1: a potent immunosuppressant. Proc Natl Acad Sci USA97: 8151–8156, 2000.

40. Yuan A, Dourado M, Butler A, Walton N, Wei A, andSalkoff L. SLO-2, a K� channel with an unusual Cl� depen-dence. Nat Neurosci 3: 771–779, 2000.

C546 PAROTID K� CHANNEL GENES

AJP-Cell Physiol • VOL 284 • FEBRUARY 2003 • www.ajpcell.org

on July 8, 2010 ajpcell.physiology.org

Dow

nloaded from